An anonymous reader shared a report. Apple is widely expected to announce major new artificial intelligence features in iOS 18 in June. The code, discovered by 9to5Mac in the first beta of iOS 17.4, shows that Apple continues to work on developing new versions of Siri that leverage large-scale language model technology, with a little help from other sources. It shows that there is. In fact, Apple appears to be using OpenAI's ChatGPT API for internal testing to help develop its own AI models. According to this code, iOS 17.4 includes a new SiriSummarization private framework that calls OpenAI's ChatGPT API. This is likely what Apple uses for internal testing of new AI features. There are also several examples of system prompts for the SiriSummarization framework in iOS 17.4. These include “Please summarize,” “Answer this question,” and “Summarize the given passage.”
It's unlikely that Apple will use OpenAI models to power iOS 18's artificial intelligence capabilities. Instead, what Apple is doing here is testing its own AI model against ChatGPT. For example, the SiriSummarization framework can perform summarization using an on-device model. Apple appears to be powering this framework with its own AI models and internally comparing its results with those of ChatGPT. In total, the iOS 17.4 code suggests that Apple is testing his four different AI models. This includes Apple's internal model called Ajax, which Bloomberg previously reported. In iOS 17.4, you can see that there are two versions of AjaxGPT, one that is processed on the device and one that is not processed on the device.
